The Founder Effect and Bottlenecks

A founder effect is the genetic consequence of establishing a new population from a small number of colonizing individuals drawn from a larger source population. The founders carry only a subset of the alleles present in the source population, and the frequencies of the alleles they do carry may differ substantially from the source purely by chance. Because the founding group is small, this is genetic drift operating at its most extreme: sampling error dominates, and allele frequencies in the new population can drift far from their ancestral values within a single generation. A population bottleneck describes the same statistical phenomenon but arising differently: an already-established population suffers a drastic, often temporary, reduction in size — from disease, habitat loss, climatic catastrophe, or overharvesting — before later recovering in numbers. The genetic diversity carried forward through the bottleneck is limited to whatever alleles happened to be present in the surviving handful of individuals. Even if the population later rebounds to millions, the diversity lost during the bottleneck generation is not restored, because population size recovery does not recreate alleles that were lost. The key quantity governing the severity of both phenomena is the effective population size, Nₑ (see Effective Population Size), not the census size. A founding or surviving group might number in the dozens, but if generations are skewed, sex ratios unequal, or variance in reproductive success high, Nₑ can be far smaller than the head count suggests. Small Nₑ means strong drift: allele frequencies bounce around unpredictably from generation to generation, heterozygosity is lost at a predictable average rate, and rare alleles are disproportionately likely to be lost entirely — or, paradoxically, driven to fixation, escaping the fate that rarity would ordinarily impose on them in a large population.

Step 1 — Sampling of gametes forms a finite zygote pool. In any sexually reproducing population, the alleles carried by the next generation are a sample of the alleles carried by gametes produced by the current generation. When population size is huge, this sampling barely perturbs allele frequencies because the sample size is enormous. When the founders or survivors number only 2–50 individuals, however, each generation's allele pool is a small random draw, and the variance of that draw (given by binomial sampling theory) is large relative to the mean.

Step 2 — Variance in allele frequency change scales inversely with Nₑ. For an allele at frequency p in generation t, the expected frequency in generation t+1 is still p, but the variance around that expectation is: Var(Δp) = p(1 − p) / (2Nₑ) Because this variance term has Nₑ in the denominator, halving the population size doubles the variance of drift — a small Nₑ makes frequencies leap unpredictably from generation to generation, in contrast to the near-constant frequencies typical of large populations.

Step 3 — Heterozygosity decays deterministically in its average rate, even though individual allele trajectories are random. Although you cannot predict which allele will be lost in any one bottleneck, the expected proportion of heterozygosity retained each generation follows a simple, testable recursion: H₁ = H₀ (1 − 1/2Nₑ) Iterated over t generations: Hₜ = H₀ (1 − 1/2Nₑ)ᵗ ≈ H₀ e^(−t/2Nₑ) A single-generation bottleneck to Nₑ = 2 (as extreme as a single breeding pair) removes on average 25% of heterozygosity in that generation alone (since 1 − 1/(2×2) = 0.75). A bottleneck to Nₑ = 500 removes only about 0.1% per generation — negligible unless sustained for many generations.

Step 4 — Rare alleles are disproportionately lost — or, occasionally, fixed. An allele present at frequency 1% in a source population of 100,000 is represented, on average, by 1,000 copies. If only 10 individuals (20 gene copies) found a new colony, the probability that the rare allele is sampled at all is small, and if sampled, it may by chance be present at a much higher frequency than 1% purely from the luck of the draw. This dual outcome — frequent total loss, occasional dramatic overrepresentation — is the mathematical signature of the founder effect and explains why rare hereditary conditions can become disproportionately common in young, small, isolated populations.

Step 5 — Recovery in numbers does not restore lost alleles. Once an allele is lost from a population (frequency reaches exactly 0), no amount of subsequent population growth can bring it back except through new mutation or migration. This means the bottleneck's genetic signature — reduced heterozygosity, skewed allele frequencies, sometimes complete loss of variation at some loci — persists indefinitely in the descendant population's genome, providing a lasting historical record of the demographic event.

Step 6 — Neutral and mildly deleterious alleles are affected the same way selection cannot discriminate at small Nₑ. Selection operates efficiently only when its strength (measured by the selection coefficient s) exceeds the strength of drift, roughly when |s| > 1/(2Nₑ) (Nₑs >> 1). In a bottlenecked population, even alleles with real, if small, fitness costs behave as if they were neutral — drift overwhelms weak selection, letting mildly deleterious recessive alleles rise to high frequency or fixation. This has direct consequences for the genetic health of bottlenecked species, foreshadowing later analysis of the Drift and Selection Interplay.

A bottleneck works identically in mechanism but differently in narrative: instead of a subset leaving to colonize, the entire population crashes and only a small remnant survives. The remnant's alleles are, again, a small random sample of what existed before the crash — but critically, this remnant is drawn from a population that may already have had substantial diversity, so the proportional loss of diversity depends on how severe and how long the bottleneck lasted (its "width" in generations as well as its "depth" in numbers).

Peter Buri's Drosophila bottle experiment (1956) is the textbook demonstration of drift's power at small Nₑ, and it is the direct experimental antecedent for founder-effect and bottleneck reasoning. Buri set up 107 separate laboratory populations of Drosophila melanogaster, each maintained at a constant Nₑ of 16 breeding adults (8 males, 8 females) per generation, all starting at the same allele frequency of 0.5 for the brown eye-color alleles (bw and bw⁷⁵, both selectively near-neutral). Because each of the 107 lines was propagated independently with the same tiny population size every generation, Buri could directly observe drift's stochastic action without confounding by migration or selection differences. Over 19 generations, allele frequencies spread out dramatically among the replicate lines: some populations fixed the bw allele completely, some fixed bw⁷⁵ completely, and the variance in frequency among lines increased almost exactly as predicted by the drift formula Var(Δp) = p(1−p)/(2Nₑ). By generation 19, 30 lines had fixed one allele and 28 had fixed the other, out of the original 107 — a direct empirical confirmation that small population size alone, with no selective advantage assumed for either allele, drives populations to fixation or loss at rates matching the theoretical prediction.

Founder events in real colonizing populations provide field confirmation. Classic studies of house sparrows (Passer domesticus) introduced to North America and silvereyes (Zosterops lateralis) colonizing Pacific islands both document rapid divergence in morphology and marker-gene frequencies among newly founded populations relative to their source, consistent with founder-driven drift rather than local adaptation, since the direction of divergence differs unpredictably among independently founded island populations.

Human founder populations offer some of the most striking natural "experiments." The Old Order Amish of Lancaster County, Pennsylvania, descend from a small number of 18th-century German and Swiss immigrant families (fewer than 200 founders for some sub-lineages). Victor McKusick's studies (beginning in the 1960s) documented that Ellis-van Creveld syndrome — an autosomal recessive disorder causing short-limbed dwarfism, polydactyly, and heart defects, caused by mutations in the EVC gene — occurs in this community at a frequency roughly 200 times higher than in the general population, traceable to a single founding couple, Samuel King and his wife, who arrived in the 1740s. Genealogical reconstruction confirmed that most affected individuals could trace both copies of the mutant allele back to this couple — a textbook founder effect, not a response to local selection pressure.

The northern elephant seal (Mirounga angustirostris) bottleneck is the classic large-mammal case. Commercial sealing in the 19th century reduced the entire species to a remnant population estimated at fewer than 30 individuals on Guadalupe Island, Mexico, by the 1890s. Protected from the 1920s onward, the population rebounded to over 200,000 individuals today. Allozyme surveys conducted by Bonnell and Selander (1974) examined 21 protein-coding loci and found zero polymorphism across the entire sampled population — every locus was monomorphic, an extraordinary result compared to related pinniped species, which typically show polymorphism at 15–20% of loci. Later mitochondrial and nuclear DNA sequencing confirmed reduced (though not literally zero) diversity, consistent with the earlier bottleneck estimate, though methodological debate continues over exactly how severe the historical bottleneck was and how much diversity has since been regenerated by mutation.

Cheetahs (Acinonyx jubatus) show an analogous, likely older and more severe, bottleneck signature. Stephen O'Brien and colleagues (1983, 1985) demonstrated that cheetahs from geographically distant African populations accept reciprocal skin grafts without immune rejection — a result normally seen only between identical twins or highly inbred laboratory mouse strains — because cheetahs are essentially monomorphic at the major histocompatibility complex (MHC) loci that normally drive rejection. Combined with very low variation at allozyme and, later, microsatellite and genome-wide loci, this pointed to one or more historical bottlenecks (proposed around the end of the last ice age, roughly 10,000–12,000 years ago) that left the species with unusually low standing genetic variation, a condition linked to elevated juvenile mortality and susceptibility to infectious disease in captive and wild populations today. Together, these lines of evidence — controlled laboratory populations (Buri), documented human genealogies (Amish), and natural demographic catastrophes with molecular confirmation (elephant seals, cheetahs) — converge on the same conclusion the theory predicts: drastic reductions in Nₑ cause disproportionate, sometimes near-total, loss of genetic diversity, independent of the fitness value of the alleles involved.

Example 1 — Heterozygosity loss after a real bottleneck (elephant seal-style calculation). Suppose a population of Nₑ = 100,000 with heterozygosity H₀ = 0.20 crashes to Nₑ = 20 for a single generation before rebounding. Using Hₜ = H₀(1 − 1/2Nₑ)ᵗ for the bottleneck generation: H₁ = 0.20 × (1 − 1/40) = 0.20 × 0.975 = 0.195 A single generation at Nₑ = 20 costs only about 2.5% of heterozygosity — not dramatic. But suppose the bottleneck persists for 10 generations at Nₑ = 20 (a more realistic scenario for a slowly recovering species): H₁₀ = 0.20 × (0.975)¹⁰ = 0.20 × 0.776 = 0.155 That is a 22% loss of heterozygosity, even though the population size itself recovers afterward — illustrating why duration of a bottleneck matters as much as its minimum size, and why elephant seals (bottlenecked for several decades, many generations, at very low Nₑ) show near-total loss at some loci while a one-generation crash would not.

Example 2 — Probability a rare allele survives a founder event. A source population has an allele A₂ at frequency q = 0.01 (so roughly 1 in 50 individuals is heterozygous, assuming Hardy-Weinberg proportions, since 2pq ≈ 0.0198). A founding group of N = 20 individuals (40 gene copies) is drawn at random to colonize a new island. The number of copies of A₂ among the 40 founder alleles follows approximately a Poisson distribution with mean λ = 40 × 0.01 = 0.4 copies. Probability of zero copies (allele lost immediately): P(0) = e^(−0.4) ≈ 0.670 Probability of at least one copy surviving the founding event: P(≥1) = 1 − 0.670 = 0.330 So roughly two-thirds of the time, this rare allele is lost the instant the colony is founded — but in the remaining one-third of cases, its frequency among founders could easily be 1/40 = 2.5%, already 2.5× its source frequency purely by chance. If, in one particular founding event, three copies happen to be sampled (P(3) = e^(−0.4)(0.4³)/3! ≈ 0.0072, a real but modest possibility), the new colony would start at q = 3/40 = 7.5%, a 7.5-fold jump from the source frequency — precisely the kind of leap that explains Amish-style disease clusters, without invoking any selective advantage for the allele.

Motoo Kimura formalized much of the mathematics that makes founder-effect and bottleneck reasoning rigorous, though the phenomena themselves were named and explored earlier by Ernst Mayr (who coined "founder effect" or "founder principle" in the 1940s–1950s while studying island bird populations) and demonstrated experimentally by Peter Buri (1956) in his Drosophila bottle-population studies. Kimura, trained originally in mathematics and cytology in Japan before working with Sewall Wright's diffusion-equation framework at Iowa State and later the University of Wisconsin, derived the diffusion approximations that describe how allele frequency distributions evolve under drift, mutation, and selection simultaneously — the same equations (Hₜ = H₀(1−1/2Nₑ)ᵗ and its extensions) used throughout this article to quantify heterozygosity loss. His 1968 proposal of the neutral theory of molecular evolution argued that most molecular substitutions are effectively neutral and fixed by drift rather than selection, a claim for which founder-effect and bottleneck dynamics provide some of the clearest supporting demonstrations, since these are settings where drift's dominance over selection is unambiguous and quantitatively predictable. Kimura's mathematical framework, alongside Buri's controlled laboratory populations and Mayr's field observations of island birds, together supply the theoretical and empirical foundation for how biologists now interpret genetic signatures of colonization and demographic crashes.

Conservation genetics depends on founder-effect and bottleneck theory to manage endangered species with small populations, such as the black-footed ferret (descended from as few as 7 founders after a 1980s captive-breeding rescue), the Florida panther (given a deliberate genetic rescue via introduction of Texas cougars in 1995 to counter inbreeding depression from a historical bottleneck), and the California condor (reduced to 22 individuals in 1987). Studbook management now explicitly tracks founder representation and mean kinship to minimize further diversity loss in captive breeding programs.

Human medical genetics uses founder-population mapping as a research tool: because founder populations like the Ashkenazi Jewish community, Finns, Icelanders, and the Amish carry a reduced number of disease-causing haplotypes at elevated frequency, geneticists can map recessive disease genes far more efficiently in these populations than in outbred ones — a strategy that contributed to the identification of BRCA1/BRCA2 founder mutations in Ashkenazi Jewish populations and numerous Finnish disease-gene discoveries.

Agricultural and biotechnology breeding programs must guard against inadvertent founder effects when establishing new breeding lines or transferring small numbers of individuals (seed stock, broodstock, or laboratory strains) between facilities, since unintentional bottlenecking can fix deleterious recessive alleles or erode traits needed for future selective breeding.

Invasive species management uses founder-effect signatures to reconstruct invasion history: low genetic diversity in an invasive population relative to its native range is itself evidence of a founder event, helping trace the geographic origin and number of introduction events, information used to design quarantine and biosecurity policy.

Ask the tutor: How would you calculate the minimum number of founders needed to retain 95% of a source population's heterozygosity after a single-generation founder event?